INTENTIONALITY ASSESSMENTS AND THE EVALUATION
OF ONLINE BEHAVIOR
By
David C. DeAndrea

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Communication
2011

ABSTRACT
INTENTIONALITY ASSESSMENTS AND THE EVALUATION
OF ONLINE BEHAVIOR
By
David C. DeAndrea
In many online settings, the content that appears on a webpage is created by both website owners
and viewers. This study employed the folk-model of intentionality to examine the process
through which people evaluate collectively created web content. Specifically, this study
investigated how website owners can alter the extent to which they are held accountable for
content posted by others by responding to the content. Results indicated that subjects perceived
website owners who responded to content generated by others as wanting and intending to share
the content to a greater extent than website owners who did not respond to the content posted by
others. In turn, perceptions of desire and intent influenced assessments of intentionality and
blame. Ultimately judgments of blame were associated with interpersonal evaluations of the
website owners. The results illustrate the explanatory utility of the folk-model of intentionality in
a new context, help clarify how components of the model interrelate, and underscore the
potential for more parsimonious models of blame.

This dissertation is dedicated to my parents, Fred and Mary Anne DeAndrea,
for all of their loving support.

iii

ACKNOWLEDGEMENTS
I will always cherish the time I spent at Michigan State University obtaining this degree.
There are many people I would like to thank for getting me here and making my experience so
memorable. First, I would like to thank my family for their unwavering support.
To my parents, I never had any doubt that you were behind me at all times. You always
promoted education and hard work. I could not have asked for better role models or parents.
Doug and Fred, it would have been easy for you to hold my superior athleticism against me.
Fortunately you took the high road, cheered me on, and supported me while I was getting my
degree. I will always be grateful for that. Janee, Maya, and Owen, you served as a source of
inspiration for me. Thank you all.
I would like to thank Joseph and Sandra Walther for everything they have done for me.
Joe was the reason I wanted to study at MSU and Sandra made sure that I came to visit. Joe, you
easily exceed every qualification there is for being a good academic mentor. You are an
unmatched scholar and an even more impressive person. Thanks for being so indispensible in so
many ways. I thank both of you for being so wonderful and will never forget your exceptional
generosity.
I was very fortunate to learn from and work with some brilliant scholars at MSU. The
members of my graduate committee are absolutely among these people. Frank, you have instilled
in me a healthy respect for science and the pursuit of knowledge. I will benefit for years to come
from your academic rigor. Bob, the same sentiments apply to you. Please keep fighting the good
fight in your neck of the woods. Tim, I could always count on you to have a well thought out
take on just about every aspect of scholarship and life in academia. Thanks for sharing your

iv

wisdom and being a good friend. I look forward to many more conversations and evenings out in
the future.
The friendships I’ve made while at MSU are a huge reason why I will always treasure my
time in East Lansing. I am very lucky to have met so many great people that I will remain friends
with for the rest of my life. Thank you all. Hillary, you are without a doubt the person most
responsible for my happiness over the last four years. You, your entire family, and all of your
friends back in Chicago have been unbelievably good to me. Thank you Gary, Suzi, Mandi, and
Randi for treating me so well.
Finally, I would like to thank all of my friends from back home, Rutgers, and UD that
helped me along the way. Lance and Andy, thanks for being all around good people. Charlie, I’m
still having huge fun. Doug and Chap, thanks for always making me feel missed when I come
home. Liam, thanks for being my partner in crime, keeping my music straight, and being so
dependable⎯Villa Hill über alles.

v

TABLE OF CONTENTS
LIST OF TABLES........................................................................................................................ vii
LIST OF FIGURES ..................................................................................................................... viii
CHAPTER 1 ....................................................................................................................................1
Intentionality Assessments and the Evaluation of Online Behavior ...................................1
Folk Model of Intentionality................................................................................................3
Technology and Impression Formation ...............................................................................4
Attributions of Intentionality for Other-Generated Content ....................................7
Comments and Intentionality...................................................................................8
Warranting Value: A Rival Hypothesis .............................................................................12
CHAPTER 2 ..................................................................................................................................14
Method ...............................................................................................................................14
Participants.............................................................................................................14
Research Design Overview....................................................................................14
Stimuli....................................................................................................................14
Procedure ...............................................................................................................16
Awareness Scale ....................................................................................................17
Desire Scale ...........................................................................................................17
Intention Scale .......................................................................................................17
Intentionality Scale ................................................................................................17
Blame Scale ...........................................................................................................17
General Perceptions of Profile Owner Scales........................................................18
CHAPTER 3 ..................................................................................................................................19
Results................................................................................................................................19
Discussion .........................................................................................................................23
APPENDICES ...............................................................................................................................35
REFERENCES ..............................................................................................................................42

vi

LIST OF TABLES
TABLE 1: MEANS AND STANDARD DEVIATIONS……………………………………… 38

vii

LIST OF FIGURES
FIGURE 1: THE FOLK MODEL OF INTENTIONALITY (MALLE & KNOBE, 1997) ..........39
FIGURE 2: SAMPLE STIMULUS (MALE, POSITIVE COMMENT, PICTURE) ....................40
FIGURE 3: SAMPLE STIMULUS (FEMALE, NEUTRAL COMMENT, STATEMENT)........41
FIGURE 4: WARRANTING VALUE OF INFORMATION .......................................................42

viii

Chapter 1
INTENTIONALITY ASSESSMENTS AND THE EVALUATION
OF ONLINE BEHAVIOR
People observe and judge the actions of others to help form interpersonal impressions.
When attempting to make sense of social behavior, perceivers pay considerable attention to the
mental states of actors (Malle, Moses, & Baldwin, 2001). Knowing or inferring others' desires
and beliefs helps structure social perception by deciphering the reasons why people act—for
what purpose, to what end. Determining the intended purpose of an action also helps in its
evaluation. It is well documented that people take into account behavioral intentions before
praising or condemning an act (Malle, 2006). Examples of how attributions of intent influence
the evaluation of behavior can be seen throughout social life (Lagnado & Channon, 2008; Malle,
2006). Given the important role judgments of intent and intentionality play in the assessment of
behavior, a systematic understanding of their inference is of value. Malle and Knobe's (1997)
folk model of intentionality explains the inference process by detailing the roles that desires,
beliefs, and intentions play in the evaluation of behavior and outlining how lay people discern
intentional from unintentional action (judgments of intentionality). The folk model of
intentionality can help clarify how people evaluate social behavior (e.g., Guglielmo, Monroe, &
Malle, 2009), yet the processes through which people form intentions to act and make
attributions of intentionality involve some complexities not explicitly addressed by the model.
Communicative behavior that occurs online via new media technology highlights such
complexities.
New, participatory technologies appearing online that feature content from website
owners and viewers can cloud judgments of intent and intentionality. For instance, Facebook

1

"friends" can write statements and post pictures on each other's profiles, essentially co-creating
self-presentations. A basic question that arises when viewing co-created websites is, who is
responsible for what? Do viewers of Facebook profiles infer that all profile content—ownergenerated or not—is intentionally communicated by the profile owner? If people viewing the
profile believe that the profile owner (1) intended to communicate all of the information, (2) was
aware of postings by others, and (3) has the skill to control what is or is not posted, then—
according to the folk model of intentionality—the entirety of the profile content is likely to be
perceived as intentionally transmitted (Malle & Knobe, 1997). Since assessments of
responsibility and blame are closely related to judgments of intentionality (Malle, 2006), it is
important to understand how people make attributions of intentionality for these kinds of
collective online behavior. This study uses the folk-model of intentionality to better understand
how people evaluate certain online behavior and uses the examination of online behavior to
illustrate how the folk-model of intentionality can be extended.
Specifically, this study has four primary goals. The first goal is to specify how new media
complicate the inference of intentionality for co-created online self-presentations. The second
goal is to identify communicative cues people utilize when making attributions of intentionality
for online behavior. The third goal is to examine how attributions of intentionality impact
evaluations of online behavior. The final goal is to refine the folk-model of intentionality by
documenting its utility in a new context and discussing the implications this study’s results have
for the model. Before tackling these issues, a review of the folk model of intentionality is
warranted.

2

The Folk Model of Intentionality
The folk model of intentionality (Malle & Knobe, 1997) breaks down the various
considerations people naturally make when evaluating the behavior of others. In particular, it
systematically outlines how people determine if an action is committed intentionally or
unintentionally (see figure 1). According to the model, social perceivers evaluate the presence of
five criteria when judging if an act was performed intentionally: desires, beliefs, intentions,
awareness, and skill. Desires and beliefs are the antecedent conditions that determine if an
intention to act is present. For instance, to conclude that a waiter intended to save a falling dish
from breaking on the floor by sliding his foot underneath it, a social perceiver would have to
infer that the waiter wanted to save the dish (desire) and thought that his foot could cushion the
impact enough to save the tumbling dish from smashing to pieces (belief). Although the
presence of a desire and belief are enough for a social perceiver to conclude that an intention to
act exists (i.e., the waiter intends to save falling dishes), their presence is not sufficient for the
determination that an act is performed intentionally.
Even when people infer that others have an intention to act, they will only conclude that
an act was performed intentionally if the awareness and skill conditions are met. The awareness
condition refers to an actor's mental state when performing a behavior: whether or not the actor
is perceived to be cognizant of what he or she is doing (Malle, 2006). The skill component refers
to an actor's perceived ability to perform an intended action successfully. Malle and Knobe
(1997) used variations of the following experimental vignette to illustrate how perceptions’ of
awareness and skill impact judgments of intentionality. Participants were informed that a man
once heard of a way to trick cashiers into giving him back too much change. When participants
were told that the man could perform the trick (skill) and he was aware he received too much

3

change (awareness), 83% of participants responded that he intentionally acquired too much
change. However, intentionality assessments were significantly lower when the man only had
awareness (26%) or skill (7%) respectively. Consistent with the folk model of intentionality,
these results illustrate how having the skill to successfully perform an action influences
judgments of intentionality. How can the man intentionally acquire too much change when he
doesn't have the requisite skill to perform the trick? Likewise, it is difficult to make attributions
of intentionality when the person performing an act is unaware of what he or she is doing. To
succinctly review, "According to the folk conception, the central antecedent of an intentional
action is an intention to perform that action, and for such an intention to be ascribed, a relevant
desire for an outcome and one or more relevant beliefs about the action leading to the outcome
are required. But in order for the action to be performed intentionally, skill and awareness have
to be present as well" (Malle, 2006, p. 66).
Technology and Impression Formation
People use new media technology to acquire information about others (Westerman, Van
Der Heide, Klein, & Walther, 2008) that has the potential to shape interpersonal impressions
(e.g., DeAndrea & Walther, 2010). A unique aspect of many emerging websites is that their
content is co-created. That is, website owners and their viewers collectively create what appears.
Although it is unclear how viewers of co-created websites delineate who is responsible for what
content and how such ascriptions impact perceptions of website owners, research has provided
insight on how other-generated content (i.e., viewer content) appearing on social network sites
impacts evaluations of the profile owner. For instance, Walther and colleagues (2008, 2009)
illustrate how comments posted by others on a Facebook profile can influence viewers' ratings of
the profile owner's physical and social attractiveness. The effect other-generated content has on

4

interpersonal impressions has been partially attributed to its "warranting" value. Warranting
value refers to the degree that viewers perceive information to be controllable by the person to
whom it refers (Walther & Parks, 2002). In the context of Facebook, the less profile content is
perceived by viewers to be under the manipulation of the profile owner, the more warranting
value it has. As such, other-generated information about the profile owner purportedly has
greater warranting value than the owner's contributions.
Although the content of social network site profiles is often co-created, the profile owner
typically has considerable control over what appears. For instance, profile owners on Facebook
have the ability to remove any unwanted content from their profile. Given this control, several
interesting questions emerge. Why would profile owners allow unfavorable other-generated
content capable of evoking negative impressions to persist on their profiles? How much
warranting value does other-generated content really have when its persistence or removal is
completely controllable by the profile owner? Does the impact of unfavorable other-generated
content differ if viewers believe profile owners did or did not intentionally share the content?
There are several reasons why profile owners might allow other-generated content that
some view negatively to persist on their profile. In general, the evaluation of behavior is a
subjective process (see Schneider, Hastorf, & Ellsworth, 1979). Information viewed negatively
by some (e.g., an employer) may be allowed to persist because it is viewed positively by others
(e.g., close friends). That people consider audience characteristics to orchestrate tailored selfpresentations emphasizes this point (Goffman, 1959; Schlenker, 1975; Tice, Butler, Muraven, &
Stilwell, 1995). Because information on websites is typically communicated in a one-to-many
fashion, the ability to target a particular audience is hampered. Ultimately, features of some
communication technology provoke multiple audience problems: "…when people find

5

themselves in the awkward position of wanting to present different impressions to two or more
people in the same social encounter" (Leary, 1995, p. 109).
A reason why other-generated content, in particular, may be allowed to persist—even
though it is viewed unfavorably by the profile owner—is because profile owners do not want to
offend the person that posted the content (Tufekci & Spence, 2007) and/or publicly dispel the
characterization the content makes. Put differently, conformity pressures can influence the selfpresentations we make (Leary, 1995). Even if other-generated content is not copasetic with one's
desired persona, social pressure may thwart its removal. Additionally, the removal of content
may discourage future contributions by users, contributions ostensibly high in warranting value.
In addition to the aforementioned reasons, profile owners may allow seemingly negative
other-generated content to persist on their profiles because it suits their interpersonal goals.
Other-generated content that makes the profile owner appear dangerous or ditzy might be
precisely what the profile owner wants. The primary goal of self-presentation is to get treated in
desired ways, which sometimes necessitates the creation of impressions that may be undesirable
to some observers but appealing to others (Leary, 1995).
Although it is possible for profile owners to share information some deem unfavorable on
their profiles intentionally, it is also possible for them to share such information unintentionally.
Consistent with models of blame (e.g., Shaver, 1985), viewers who believe a profile owner
intentionally shared negative other-generated content—such as unflattering photographs or
offensive comments—should be viewed more critically than a profile owner who never wanted
or decided to share the same content. Yet, due to features of new communication technology, the
inference of intentions and intentionality can take on additional complexities.

6

Attributions of Intentionality for Other-Generated Content
The folk model of intentionality provides a basis for examining these complexities and
understanding how judgments of intent and intentionality are made in online contexts where web
content is collectively created. The connection between the act (sharing a picture) and the actor
(site owner or viewer) is blurred when website content is co-created. Recall the typical hierarchy
of intentional action: a desire (wanting to appear a certain way) and a belief (this picture depicts
me that way) produce an intention to act (deciding to share the picture with others). Actors must
be aware of what they are doing (know they are sharing a picture) for social perceivers to infer
that they performed an act intentionally. Paradoxically, with other-generated content, site owners
are first made aware that they are sharing new information on their website (either by viewing
the website/profile or receiving a notification that content as been added) and subsequently must
decide if they want to continue sharing the information with others.
For instance, social network site users can hope or wish (desire) that people in their
network post flattering pictures or comments on their profile, however, they cannot decide or
choose (intend) for them to do so. The content of an intention is always an action performed by
the person holding the intention—people cannot choose or decide the behavior of others (Baier,
1970; Malle & Knobe, 2001). Once friends do post content on a social network site profile, in
many cases, the owner can decide to share the information with others or remove it. In turn,
inferences can be made by subsequent viewers as to whether the information is shared
intentionally by the profile owner.
Unlike forwarding an email, intentionally sharing other-generated content online often
does not require an additional act. The removal of other-generated content is a seemingly
straightforward indicator that the profile owner chose not to share information posted by others.

7

However, is the converse as clear-cut? Does the persistence of other-generated content signify
that the profile owner intends to share the information and is doing so intentionally?
Clearly, profile owners cannot intend to share other-generated content if they are not
aware it exists. How then does the viewer of a Facebook profile infer that the profile owner is
aware of content posted by others? If the viewer of a Facebook profile infers that the profile
owner is aware of other-generated content, does that signify to the viewer that the profile owner
intends to share the information as well? The following section explores these questions and the
communicative cues found on co-created websites that may inform judgments of intentionality.
Comments and Intentionality
Computer-mediated communication research has shown that people are adept at using
online cues to guide inference and impression formation when nonverbal cues are absent
(Walther, 1996, Walther & Parks, 2002). On many co-created websites, website owners can
remark or "comment" on the content that others have posted. Website owners that append a
comment to the postings of others clearly signal their awareness of the other-generated content.
For instance, to append a comment to other-generated content on Facebook, profile owners must
first recognize/locate the posting, click the "comment" link below the posting, enter a comment,
and finally click the "comment" submission button. Given that other Facebook users are aware of
this process, profile owner comments on other-generated content provide a clear cue of owner
awareness. Regardless of what profile owner comments say, they unequivocally indicate that
profile owners are not oblivious to the postings of others.
H1: Profile viewers are more likely to infer that profile owners are aware of othergenerated content when it is accompanied by a profile owner comment relative to when it
is not accompanied by a comment.

8

Although perceptions of awareness play a central role in judgments of intentionality, an
action is unlikely to be viewed as intentionally performed in the absence of an intention (Malle &
Knobe, 1997). Once profile owners are aware that they are sharing other-generated content, they
are afforded the opportunity to make a decision about its persistence or removal. As such, profile
owner awareness might signal to viewers that the owner decided not to remove the content. That
is, once profile owners are perceived to be aware that they are hosting other-generated content,
the failure to remove the content serves as an indicator of tacit approval.
Silence and the failure to act can serve as the basis for a variety of inferences. For
instance, silence in response to a criminal allegation is admissible evidence of guilt in the court
of law (Tiersma, 1995) and silence can lead to false perceptions of consensus surrounding public
opinion (Noelle-Neumann, 1974). Notably in interpersonal communication, silence can serve a
judgmental function by indicating approval or disapproval (Jensen, 1973; Johannsen, 1974). If
the persistence or removal of other-generated content is used by profile viewers to gauge whether
or not profile owners intend to share the content, it would be important for viewers to rule out the
possibility that the content persists because the profile owner is unaware of its existence.
Commenting on other-generated content establishes that profile owners are aware that the
content exists. The failure to remove such content can be used by viewers to infer that profile
owners chose (intended) to share it with others. When profile owners are aware of othergenerated content, they have the opportunity to remove it, whereas profile owners unaware of
other-generated content never have the opportunity to choose whether the content should persist
or be removed.

9

H2: Profile owners' comments indirectly influence viewers' perceptions of whether
profile owners intend to share other-generated content through viewers' perceptions of
profile owner awareness.
In addition to influencing perceptions of whether profile owners are afforded the choice
of sharing other-generated content, some comments by profile owners may provide direct
evidence about their desire to share the information. Beyond signaling their awareness,
comments can indicate how profile owners feel about other-generated content being shared.
Viewers can infer from a comment that favorably responds to a picture or message posted by a
friend that the profile owner wants to continue sharing the content with others, whereas a neutral,
non-evaluative comment may only indicate awareness. Malle and Knobe (1997) provide
evidence that inferences of intent (i.e., an actor choosing a course of action [e.g., choosing to
share a picture on Facebook]) are significantly impacted by the presence/absence of a desire.
H3: Profile viewers perceive profile owners as having a greater desire to share othergenerated content when profile owners append a favorable comment relative to a neutral
comment.
It has been argued that profile owner comments appended to other-generated content can
function to influence the presence of three critical criteria perceivers rely on to determine if an
act was performed intentionally: desires, intentions, and awareness. These criteria have been
documented (Malle & Knobe, 1997) to significantly influence judgments of intentionality (i.e.,
an actor choosing a course of action and carrying out the act with skill and awareness [e.g.,
intentionally sharing a picture on Facebook]). Viewers would be hard-pressed to infer that profile
owners are unaware of other-generated content that they directly commented on. If viewers infer
that profile owners are aware of other-generated content, they may also infer that profile owners

10

had an opportunity to remove the content. Beyond exclusively indicating awareness, the
substance of profile owner comments may indicate to viewers whether or not profile owners
desire to share other-generated content on their profiles. Accordingly, the following hypothesis is
offered:
H4: The influence of profile owner comments on viewer perceptions that other-generated
content is intentionally shared by profile owners follows an ordered monotonic trend such
that favorable comments lead to the greatest perceptions of intentionality, followed by
neutral-valenced comments, followed by no comment.
In general, the degree to which a person is blamed or praised for an act is largely
contingent on whether the behavior is perceived to have been performed intentionally or
unintentionally (Malle, Moses, & Baldwin, 2001). People try to disassociate themselves from
their negative actions in order to preserve desired social identities, understanding that the
magnitude of social sanctions partially depends on whether the transgression is believed to have
been committed intentionally or accidentally (Schlenker & Weigold, 1992).
"…it is clear that the greatest amount of either causality or blameworthiness will be
attributed to a stimulus person who has intentionally produced a negative outcome"
(Shaver, 1985, p. 96).
Empirical evidence supports the claim that actors receive greater blame when negative acts are
perceived to have been intentional rather than unintentional (Lagnado & Channon, 2008; Schultz
& Wright, 1985).
H5: Intentionally shared other-generated content that is negative leads to more blame of
the profile owner for sharing the content than unintentionally shared other-generated
content that is negative.

11

Ultimately, exploring how people make intentionality assessments is important for this very
reason. In an online context, viewers who blame profile owners for sharing negative othergenerated content should judge profile owners more harshly than viewers who do not believe the
profile owners should be held accountable for the negative content.
H6: The more viewers blame profile owners for sharing negative other-generated content,
the less favorably profile owners are evaluated.
Warranting Value: A Rival Hypothesis
Previous research suggests that the effect online information has on interpersonal
impressions is partially a function of its warranting value, the degree that viewers perceive
information to be uncontrollable by the person to whom it refers (Walther & Parks, 2002). The
folk model of intentionality and models of blame are not wholly consistent with previous
warranting research. The warranting value of online information has been experimentally varied
by creating mock Facebook profiles with owner-generated and other-generated content where
other-generated content was used to represent information not controllable by the profile owner
(Walther et al., 2009). However, as previously suggested, other-generated content is controllable
with regard to its persistence on Facebook and other social network sites. The extent to which
owner-generated and other-generated content is perceived to be controllable by the profile owner
has not been directly measured in past research. Rather, who ostensibly produced the content
served as the basis for conceptualizing and testing a warranting effect.
According to the folk-model of intentionality (Malle & Knobe, 1997), awareness is a
necessary component for attributions of intentionality, which in turn impact assessments of
blame (Lagnado & Channon, 2008). Measuring whether profile viewers think profile owners are
aware of other-generated content can be used to test the folk model of intentionality and the

12

purported link between intentionality and blame assessments. The assessment of profile owner
awareness, however, also permits a test of the warranting hypothesis. Although profile owners
often have the capacity to remove other-generated content from their profiles, they cannot
remove content they do not know exists. By definition, other-generated content should have
greater warranting value when viewers believe that profile owners cannot control its
presence/removal. The warranting hypothesis suggests that content high in warranting value
(uncontrollable) has a greater impact on interpersonal evaluations than content low in warranting
value—such as controllable other-generated content that profile owners are aware of and could
remove. As such, the folk model of intentionality offers a rival hypothesis (see H) to that of
warranting theory, where warranting offers the following prediction:
H7: Other-generated content that is negative leads to more unfavorable impressions of a
profile owner the less viewers believe the owner is aware of its presence.

13

Chapter 2
Method
Participants
A total of 306 undergraduate students were recruited and offered extra credit or partial
fulfillment of a course research requirement in return for their participation.
Research Design Overview
Mock Facebook profiles were created and served as the stimuli for this study. Negative
other-generated content, ostensibly posted by a Facebook friend, appeared in all conditions.
Three factors were experimentally varied across the stimuli to examine the hypotheses: the
profile owner’s comment/reply to the other-generated content, the profile owner’s sex, and the
type of negative other-generated content appearing on the mock profile.
Profile owner comment/reply type was a between-subjects factor with three conditions:
favorable comment, neutral comment, and no comment. The gender of the profile owner also
served as a between-subjects factor. That is, there was a male and female version of all mock
profiles. This allowed for any gender effects to be detectable. The final factor was the type of
other-generated posting. This between-subjects factor contained two conditions: written
statements and pictures. Research by Walther et al. (2008) indicates that both types of postings
can affect interpersonal perceptions via Facebook. In sum, the study employed a three (comment
type) by two (gender of profile owner) by two (posting type) crossed experimental design.
Stimuli
As noted, mock Facebook profiles served as the stimuli for this study. Since the type of
negative, other-generated content served as an experimental factor, mock profiles were created
that contained either a written message or a picture. For the written message condition, a

14

statement appearing as a wall posting from a Facebook friend read, "You were definitely the
most obnoxious drunk at Ricks last night. Do you even remember getting thrown out?" For the
picture condition, a photograph depicting the profile owner simultaneously drinking two beers
appeared to be posted by a Facebook friend.
As noted, the type of comment or reply the profile owner made toward the othergenerated content was an experimental factor with three conditions. For the favorable response
condition, the profile owner appeared to have appended the following comment, "ha, typical
Friday night.” For the neutral response condition, the profile owner appeared to have appended
the following comment, “You around this weekend?" A no comment control served as the third
condition. The creation of these stimuli was guided by the results of two pilot tests.
The first pilot study was conducted to determine what type of Facebook content people
view negatively. Participants included 179 students from a subject pool that approximates the
one used in this study. After completing a questionnaire from a separate study, participants were
directed to answer the following: "Please provide an example of a picture that others could
post/tag of you on Facebook that you would want to delete" and "Please provide an example of a
picture you have viewed on Facebook that made you form a negative impression of the person in
the photograph." For the first question, 50% of respondents mentioned pictures of them drinking
or intoxicated, whereas 45% of respondents mentioned pictures of a person drinking or
intoxicated for the second question⎯each being the modal response. Accordingly, stimuli
depicting the profile owner as having been in an intoxicated state were used for the study.
Since the gender of the profile owner varied across conditions, a second pilot test was
conducted to determine if the male and female depicted in the pictures varied in attractiveness or
the pictures varied in positivity/negativity. A total of 69 participants were recruited from a

15

separate group of subjects, similar to the participants used in the main study. The physical
attractiveness of the male and female target was rated on a 5-point scale, with endpoints ranging
from “very unattractive” to “very attractive.” The attractiveness of the male and female targets
did not significantly differ, t (65) = 1.18, p = .24. Subjects also evaluated how positively,
negatively, favorably, and unfavorably (α = .90) they viewed the targets based on how they were
depicted in their respective photographs (7-point scales; higher scores indicate positive
perceptions). The male and female targets were, on average, viewed negatively, M = 3.17, SD =
1.19, and not significantly different, t (67) = .348, p = .73 (see Figures 2 and 3 for sample
stimuli).
Procedure
Subjects were led to a small, private research room containing a computer and given an
informed consent form. After providing consent, subjects were randomly assigned to view one
of the mock Facebook profiles. They were told that their task was to view content from a
Facebook profile and answer questions about it. Participants were informed that the researchers
were given permission from some students at the University to monitor and capture screen shots
from their Facebook profiles. Although the researchers were given permission to monitor the
Facebook profiles of some students over the course of a month, it was impossible to get
permission from all of the students’ Facebook friends. That is why all photographs identifying
Facebook friends were blurred out (ostensibly). In reality, the same blurred friend photo was
used in every condition to control possible extraneous effects. After viewing the stimulus,
participants completed an online questionnaire. Finally, the subjects were thanked and debriefed.

16

Awareness Scale
A four item measure was created for this study to assess the likelihood that the profile
owner was aware of other-generated content on his/her Facebook profile (see Appendix A). All
items in this study were measured on 7-point scales with endpoints ranging from "strongly
disagree" to "strongly agree." Conceptual definitions from Malle and Knobe (1997) were used to
guide the creation of the items. Reliability was assessed via Cronbach’s alpha (α = .95).
Desire Scale
A four item measure was created for this study to assess perceptions that profile owners
had a desire to share the other-generated content on their profile (see Appendix A). The items
contained verbs from Malle and Knobe (2001) that reflect desire (α = .92).
Intention Scale
A four item measure was created for this study to assess perceptions that profile owners
intended to share other-generated content on their Facebook profile (see Appendix A). The scale
utilized verbs reflecting "intention" in Malle and Knobe (2001; α = .86).
Intentionality Scale
A four item measure was created for this study to assess perceptions that the profile
owners intentionally shared other-generated content on their Facebook profile. The scale creation
was informed by Malle and Knobe (1997; see Appendix A). An item was dropped to increase the
reliability of the scale (α = .76)
Blame Scale
A four item measure was created for this study to assess perceptions that the profile
owner was to blame for sharing the other-generated content (see Appendix B). One item was
dropped to increase the reliability of the measure (α = .76).

17

Perceptions of Profile Owner
Two scales were used to assess participants’ perceptions of the profile owners. A four
item measure was adapted from DeAndrea and Walther (2010) that assessed a general
favorability rating of the Facebook profile owner (see Appendix B; α = .93). A three item
measure was created to assess academic task attraction from more general items provided by
McCroskey and McCain (1974). All items appear in Appendix B (α = .93).
The internal consistency of each scale with at least four items was assessed using Hunter
and Hamilton’s (1992) Confirmatory Factor Analysis Program. The program provides factor
loadings that were used to calculate predicted inter-item correlations. For each scale, deviations
were computed between predicted and obtained inter-item correlations. The root mean square
error was then calculated for each scale: awareness (.03), desire (.03), intent (.09), and
favorability (.04).

18

Chapter 3
Results
When profile owners respond to [negative] content others post to their webspace, how are
viewers’ perceptions of the content and the profile owner affected? The current study addressed
this question by examining how profile owner comments affect components of the folk model of
intentionality (i.e., awareness, desire, intent), and in turn, evaluations of online behavior. To
increase the generalizabilty of the findings, the type of negative other-generated content
(statement or picture) and the sex of the target being evaluated (male or female) were also varied.
The sex of the target did not moderate the effects of profile owner comments (favorable, neutral,
control) on any of the dependent measures, nor did it exert any main effects. The type of othergenerated content (statement or picture) did, however, affect perceptions of desire and
intentionality. As such, the following analyses report the effects of profile owner comments on
the dependent variables across these two factors—with an exception for analyses involving the
dependent measures of desire and intentionality.
The presence/absence of profile owner comments is expected to increase viewers’
perceptions that profile owners are aware of other-generated content (H1). This hypothesis was
examined using an independent samples t-test to compare the neutral comment condition, M =
6.30, SD = 1.30 with the control comment condition, M = 4.45, SD = 1.60. Results indicate that
viewers perceive profile owners to be aware of other-generated content to a significantly greater
2

extent when a comment appears, t (202) = 9.03, p < .01, ω = .28.
The presence/absence of profile owner comments is predicted to increase viewers’
perceptions that profile owners intend to share other-generated content (H2). This hypothesis
was examined using an independent samples t-test to compare the neutral comment condition, M

19

= 4.52, SD = 1.38, with the control comment condition, M = 3.14, SD = 1.37. Results indicate
that viewers perceive profile owners as intending to share other-generated content to a
2

significantly greater extent when a comment appears, t (201) = 7.13, p < .01, ω = .20.
Because the effect of profile owner comments on perceptions of intent was argued to
occur indirectly through perceptions of awareness, an additional analysis was employed to more
precisely examine H2. A Sobel test was conducted to directly examine if the total effect of
comment presence/absence (X) on perceptions of intent (Y) is significantly reduced when
awareness is accounted for as a mediator (M). Using the macros provided by Preacher and Hayes
(2004), a significant indirect effect was detected, z = 5.10, p < .01, rXY = .45, rXY.M = .22.
Hypothesis 3 specifies that, relative to neutral profile owner comments, favorable
comments increase viewers’ perceptions that profile owners desire to share other-generated
content. This hypothesis was examined using an independent samples t-test to compare the
favorable comment condition, M = 5.61, SD = 1.11, with the neutral comment condition, M =
4.93, SD = 1.40. Results indicate that viewers perceive profile owners as having a significantly
greater desire to share other-generated content when it is accompanied by a favorable comment
2

relative to a neutral comment, t (201) = 3.89, p < .01, ω = .07. Additionally, a two-way analysis
of variance (ANOVA) indicated that posting type (statement/picture) interacted with comment
2

type (favorable/neutral) to affect perceptions of desire, F (1, 199) = 4.30, p = .04, ω p = .02. The
type of posting did not affect perceptions of desire across the favorable comment conditions, t
(100) = .04, p = .97, whereas it did significantly affect perceptions of desire across the neutral
comment conditions, t (100) = 2.65, p = .01. Table 1 provides means and standard deviations.

20

As stated in hypothesis 4, profile owner comments are predicted to affect viewer
perceptions that other-generated content is intentionally shared in the following manner:
favorable comments lead to the greatest perceptions of intentionality, followed by neutral
comments, followed by no comment. The hypothesized trend (H4) was examined via a contrast
analysis. The following contrast coefficients were assigned to the three conditions: favorable
comment (+2), neutral comment (+1), and no comment (-3). The selection of unequal-interval
coefficients was guided by theory: favorable comments were expected to satisfy three criteria of
intentionality (i.e., desire, intent, awareness), neutral comments were expected to satisfy two
criteria (awareness and intent), whereas the control condition was not expected to explicitly
satisfy any components identified in the folk-model of intentionality. Overall, the data were
2

consistent with the hypothesized pattern of means, t (302) = 6.38, p < .01, ω φ = .12. Perceptions
of intentionality were highest in the favorable comment condition, M = 4.64, SD = 1.46,
followed by the neutral comment condition, M = 4.46, SD = 1.40, followed by the no comment
control condition, M = 3.50, SD = 1.25.
Although posting type (statement/picture) interacted with comment type to affect
2

perceptions of intentionality, F (2, 299) = 4.64, p = .01, ω p = .02, the hypothesized pattern of
2

means remained significant across the statement, t (149) = 5.64, p < .01, ω φ = .17, and picture
2

conditions, t (150) = 3.60, p < .01, ω φ = .07. The favorable and neutral comment conditions
were always significantly greater in perceived intentionality than their respective control
condition (all p’s < .01; see Table 1). The significant comment type by posting type interaction
is attributable to the following difference. When the other-generated content was a statement,
perceived intentionality was significantly greater in the favorable comment condition relative to

21

the neutral comment condition, t (98) = 2.75, p < .01. When the other-generated content was a
picture, perceived intentionality did not significantly differ across the favorable and neutral
comment conditions, t (101) = 1.36, p = .18. Table 1 provides means and standard deviations.
Viewers’ perceptions that a profile owner intentionally shared other-generated content are
expected to be positively associated with perceptions that the profile owner is to blame for
sharing the other-generated content (H5). A significant, positive association was found between
the two variables, r (305) = .46, p < .01. The results of the tests for hypothesis 4 and 5 indicate
that comments influence perceptions of intentionality, and in turn, perceptions of intentionality
are associated with judgments of blame. The following tests were conducted to more closely
examine this process.
Theoretically, commenting on other-generated content should impact perceptions of
blame through perceptions of intentionality. Accordingly, the same contrast coefficients
employed to test the effect of comments on perceptions of intentionality were used to test the
effect of comments on perceptions of blame. The data were consistent with the predicted pattern,
2

t (302) = 4.82, p < .01, ω φ = .07. Perceptions of blame were greatest in the favorable comment
condition, M = 5.42, SD = 1.20, followed by the neutral comment condition, M = 5.10, SD =
1.41, followed by the control condition, M = 4.47, SD = 1.61. The mere presence of a comment
(neutral comment vs. no comment) led to greater judgments of blame, t (201) = 2.95, p < .01, ω
= .04, and a Sobel test indicated a significant indirect effect of comment presence (X) on
perceptions of blame (Y), mediated through perceptions of intentionality (M), z = 3.83, p < .01,
rXY = .20, rXY.M = .07.
Viewers’ perceptions that a profile owner is to blame for sharing [negative] othergenerated content is anticipated to be negatively associated with viewers’ interpersonal
22

2

evaluations of the profile owner (H6). That is, the more viewers blame a profile owner for
sharing negative information online, the less favorably they evaluate the profile owner. Two
bivariate correlations indicated that the data were consistent with this expectation. Viewers’
perceptions of blame were negatively associated with their general evaluation of the profile
owner, r (305) = -.12, p = .038, and their academic task attraction to the profile owner, r (305) =
-.16, p < .01. Specifically, the more viewers blame a profile owner for sharing negative othergenerated content, the less favorably they view the profile owner and the less willing they are to
rely on the profile owner to support their academic pursuits.
A rival hypothesis (H7) derived from research on the warranting value of online
information predicted that the less viewers perceived profile owners to be aware of othergenerated content (i.e., the more uncontrollable the content), the more negatively viewers would
evaluate the profile owner. Bivariate correlations indicate that the data were not consistent with
this prediction. Perceptions of awareness were not associated with assessments of favorability, r
(305) = -.04, p = .48, or assessments of academic task attraction, r (305) = -.10, p = .075.
Discussion
The results of this study document the utility of the folk-model of intentionality for
understanding how people evaluate certain online behavior. At the same time, the collective
nature of the online behavior examined helps highlight areas where the folk-model of
intentionality can be extended. A brief review of the findings precedes a detailed discussion of
these claims.
The results of this study demonstrate that viewers of websites use profile owner
comments to determine if profile owners are aware of content posted by others. In turn, the more
viewers believe that profile owners are aware of content posted by others, the more likely they

23

are to believe that the profile owners intended to share the content. When profile owners do
make comments, favorable responses increase viewers’ perceptions that the profile owner wants
to the share the content, relative to neutral, non-evaluative replies. Consistent with the folkmodel of intentionality, perceptions of awareness, desire, and intent influence whether or not
viewers’ think profile owners intentionally share content posted by others. Notably, the more
viewers think that profile owners are intentionally sharing other-generated content, the more they
blame the profile owner for sharing the content. Finally, perceptions of blame are associated with
harsher interpersonal evaluations of profile owners.
An abundance of social network sites, personal blogs, and other websites are currently
being used by millions of people to interact with family, friends, acquaintances, and co-workers.
People have an inherent drive to manage how they are viewed during social encounters
(Goffman, 1959) and to use information from social encounters to form impressions of others
(Malle, 2004). These basic motives, manifested through communication, are not vanquished
because social interactions occur online. Website owners can greatly influence the extent to
which they are held accountable for the contributions of others by simply acknowledging their
awareness of them. Postings on Facebook and other social network sites have reportedly cost
people their jobs (Valle, 2008), their marriages, and custody of their children (Davis, 2010). In
such scenarios, attributions of what content people are and are not responsible for on their own
profile can take on considerable moral and legal weight.
This study illustrates how the folk-model of intentionality can be applied to online
settings to help understand how people evaluate mediated communication. Importantly, when
profile owners respond to or comment on the postings of others, viewers are provided a
considerable amount of information. Communicative cues, in the form of comments, indicated if

24

a profile owner wanted to, intended to, and purposefully shared content posted by others. Theory
(Guglielmo et al., 2009; Shaver, 1985) and empirical findings (Lagnado & Channon, 2008;
Malle & Knobe, 1997) indicate that these considerations play a critical role in attributions of
blame. The results of this study support the claim that intentionality assessments are related to
attributions of blame and help explain when negative, other-generated content is most likely to
damage viewers’ perceptions of profile owners. In doing so, insight is provided for individuals
who wish to manage their impressions in online settings. Profile owners can promote positive
impressions by strategically reacting to the content others post.
It is possible that some social perceivers may not take intentions into account when
viewing negative other-generated content online. For instance, an employer may not care if a
potential employee did not want to share a photograph that a friend posted of the prospective
employee drunk. Employers may simply use information found online to guide hiring decisions,
regardless of whether the information was shared intentionally. In general, other-generated
content viewed negatively may produce unfavorable reactions regardless of judgments of intent.
Alicke (2008) notes that people are quite capable of spontaneously evaluating behavior in ways
wholly inconsistent with rational models of blame and decision making. Yet there is ample
evidence that perceptions of intent and intentionality often play a role when people assess blame.
Although this study provides considerable insight into how comments affect viewers’
perceptions of other-generated content, it does not address what occurs when profile owners
respond unfavorably to others’ postings. That is, this study only examined the effect of positive
and neutral comments. An unfavorable response appended to other-generated content by a
profile owner may indicate little desire to share the content. On the other hand, because the
other-generated content is allowed to persist, it may ultimately indicate some desire to share the

25

content (i.e., actions speak louder than words). For this reason, it is unclear if unfavorable or
contradictory comments would lead to greater/lesser attributions of intentionality and blame than
no comment. Understanding how viewers perceive the appropriateness of deleting content posted
by others may help clarify this dilemma. If viewers expect profile owners to delete othergenerated content that they deem inappropriate, the persistence of such content may be viewed as
an indicator of approval. A negative response to the content might not be seen as sincere or a
strong enough rebuke of the content. In contrast, if viewers believe that the social costs of
deleting content posted by others outweighs what might occur by allowing the content to persist,
a negative response may be viewed as an appropriate and sincere indicator of disapproval.
Viewers’ perceptions of how much a profile owner should monitor their profile and delete
content posted by others needs to be taken into account in future research that examines how
collective online behavior is evaluated. In addition to examining the effect unfavorable or
disconfirming comments have on perceptions of other-generated content, future research might
benefit from examining the perceived relationship between website owners and individuals
posting content.
Other-generated content may reflect differently upon a profile owner when posters are
viewed as friends of the profile owner rather than strangers with no clear relation to the profile
owner. For instance, inflammatory remarks written on a Facebook wall may reflect strongly on
the profile owner, whereas the same inflammatory remarks may not have the same effect if
written in response to a blog post. Are there normative perceptions associated with different
online media about the relationship between profile owners and viewers? Research suggests that
Facebook friends are often not intimately familiar with one another (Parks, 2010). Nevertheless,
viewers might not be aware of this reality and they might not be able to discriminate between

26

actual friends and loose acquaintances. Are there communicative cues that indicate whether or
not an owner has a relationship with a given comment poster? Empirical investigations are
needed to answer these questions.
Further research is also needed to explore more closely the warranting hypothesis in
online and offline settings. At the core of the warranting hypothesis is a prediction about the
perceived controllability of information. People can control what they say about themselves to a
greater extent than what others say about them. Because social perceivers recognize this, they
rely more heavily on what others have to say about an individual to guide their impressions
(Walther & Parks, 1992). After all, people are usually motivated to present themselves in a
favorable light. This logic, applied to information posted on Facebook, predicts that othergenerated content about profile owners will impact viewers’ perceptions more heavily than
content profile owners provide about themselves (Walther et al., 2009).
It was argued in this manuscript that previous warranting research (e.g., Walther et al.,
2009) did not provide a robust examination of the warranting hypothesis because the perceived
controllability of information was not measured. That is, an induction check of the experimental
manipulation was not performed. Rather, it was assumed that by varying the source of the
information (profile owner or Facebook friend), perceptions of the controllability of the
information varied as well⎯profile owners would be seen as having less control over othergenerated content than self-generated content. On Facebook, however, profile owners do have
extensive control over what appears regardless of who generated the content. This reality
suggests the need for a more nuanced conceptualization and measurement of the warranting
value of information.

27

Control over the generation and dissemination of content should be considered. For
instance, consider the warranting value of information about a dictator. The dictator says he is a
munificent ruler beloved by his people. The dictator’s national media broadcasts also say that he
is a wonderful leader. Is the warranting value of this information really that dissimilar? In a
sense, Facebook users are dictators of their own online fiefdoms. They control who can see their
profile, who can post content to their profile, and what content can persist on their profile.
Facebook users may be much more inclined to write something negative about a “friend” than
people living under the rule of a dictator would dare write about their supreme ruler. But profile
owners on Facebook do have considerable control over what information is disseminated to the
masses. When the dissemination of other-generated content is completely controllable, its
warranting value might not be much different than self-generated content.
This study took a different approach to assessing the warranting value of online
information. It was argued that if viewers believe profile owners are aware of other-generated
content, they are likely to believe that profile owners have control over the content. If viewers
believe profile owners are unaware of other-generated content, they are likely to believe that
profile owners do not [currently] have control over the content. As such, perceptions of
awareness serve as a direct indicator of the controllability of information presented on Facebook.
Consistent with the warranting hypothesis, negative other-generated content would carry more
weight when viewers thought profile owners were unaware of the content relative to when they
thought profile owners were aware of the content. The data were not consistent with this
prediction. Perceptions of awareness were not associated with interpersonal evaluations of the
profile owner.

28

Although these results raise challenges for the warranting hypothesis, the methodology
employed in this study suffers from its own limitations. Notably, the source of the content was
held constant. Only control over the dissemination, not the generation, of the information was
varied. A more complete examination of the warranting hypothesis is needed that varies who
generates information (self vs. other) and the extent to which the source has control over its
dissemination. Importantly, perceptions of controllability should be measured directly.
Future studies that examine the warranting hypothesis should also consider the
implications of previous impression management research. For instance, research suggests that it
is can be advantageous for people to divulge negative information about themselves before
others have a chance to do so. Individuals and organizations that have committed transgressions
have been able to use the impression management tactic of “stealing thunder” (Williams,
Bourgeois, & Croyle, 1993) to reduce negative evaluations and sanctions (Arpan & RoskosEwoldsen, 2005). By revealing negative self-information before others, people are able to
increase their credibility and more favorably frame the negative event (Arpan & RoskosEwoldsen, 2005). In this study, another person always disseminated the negative information
first, eliminating the possibility of a stealing thunder effect. Warranting research that examines
the implications of competing sources of negative information (self/other) should consider how
the information is temporally distributed.
Before shifting focus to what this study entails for the folk-model of intentionality and
models of blame, a brief discussion of the effects of posting type (picture or statement) is needed.
First, it is worth noting that a picture can be more flattering, damaging, or mundane than a
statement and vice versa. It has been demonstrated that pictures can have a stronger effect than
statements on interpersonal evaluations in online settings (Van Der Heide, D’Angelo, &

29

Schumaker, in press). However, the ubiquity of such effects and the causal mechanism driving
such effects is unclear. A directional hypothesis predicting that statements have a stronger effect
than pictures on interpersonal evaluations would have been supported in this study. Subjects
viewed profile owners less favorability in the written statement conditions, M = 3.36, SD = 1.26,
relative to the picture conditions, M = 3.63, SD = 1.23, t (303) = 1.89, p = .03, 1-tailed. The
pictures and statements in this study were designed to elicit negative impressions, however, it is
unclear if they were equally negative. As stated in the results, viewers perceived profile owners
as wanting (desire) to share the pictures to a greater extent than the statements. This would make
sense if the subjects felt that the pictures were less damning than the statements. Overall, it is
important to consider that many features/characteristics are nested within types of information
(e.g., pictures, written statements, oral claims).
Future investigations may benefit by focusing more on differences in information than
differences in information type. For instance, studies that examine the verifiability or vividness
of other-generated content might be more informative than studies that look at differences
between pictures and statements per se. In this study, it was relatively clear that the profile
owners were drinking alcohol in the picture conditions. In the statement conditions, there is only
an accusation of drunken behavior by a friend. Nevertheless, the unverified statements led to
predominantly negative judgments of the profile owners. This speaks to the power others can
have over how people are viewed online. The verifiability of each type of information (picture
vs. statement) was confounded with its positivity/negativity in this study but need not be in
future research.
Finally, this study’s findings have important implications for the folk-model of
intentionality and models of blame. Although previous research has detailed how components of

30

the model individually affect intentionality assessments, the connections between all of the
model’s components have not been fully examined. This study illustrates how perceptions of
awareness can influence perceptions of intent and desire. These possibilities have not been given
extensive consideration previously because desires are considered antecedents of intentions and
intentions are considered antecedents of action. As such, both desires and intentions precede
conscious awareness that an act is taking place. But it is important to keep in mind that
intentionality assessments are post-hoc evaluations of behavior (Alicke, 2008). Also, people can
be judged not only if they intentionally performed an action, but also if they intentionally
supported or permitted an act to occur or persist.
It would not make sense to claim that a profile owner intentionally posted content
generated by others. People can manipulate, persuade, and threaten others, but they cannot
choose the actions of others and they cannot intentionally carry out others’ actions. People can,
however, intentionally allow actions to occur or persist. In such scenarios, the components of the
folk-model of intentionality can still be used to explain the inference process. When an
individual does not initiate an action, awareness that the act is occurring takes on a new role of
considerable importance. As illustrated in this study, viewers use perceptions of awareness to
determine that an actor had an opportunity to decide whether to allow an action to continue.
Furthermore, the failure to halt an act appears to indicate tacit approval and a desire for the act to
persist. As anticipated by the folk-model of intentionality, perceptions of awareness, intent, and
desire impacted intentionality assessments, which affected attributions of blame. The more
viewers blamed the profile owners for sharing the information, the more negatively they
evaluated the profile owners. By showing how considerations of awareness can influence
perceptions of desire and intent, and ultimately, intentionality and blame, this study reveals a

31

new application of the folk-model of intentionality and a different understanding of how its
components interrelate.
By expanding the applicability of the folk-model of intentionality, operational
redundancies that exist in models of blame may be eliminated. For instance, the model presented
by Guglielmo et al. (2009) identifies distinct routes leading to blame that hinge on assessments
of intentionality. That is, both intentional and unintentional actions can lead to blame, but the
paths culminating in judgments of blame differ. As emphasized by Hamilton (1978), in order to
fully understand attributions of responsibility and blame, it is important to consider what should
have been done in addition to what was done. Indeed, "…liability holds even for unintentionally
caused outcomes, as long as the agent had the intentional capacity and duty to prevent that
outcome" (Malle, Moses, & Baldwin, 2001, p. 21). After perceivers determine that an act or
event was unintentional, they can still blame actors if they believe that the actors could have
foreseen or prevented the event.
Consistent with this claim, Fincham and Jaspers (1983) found that the foreseeability of a
negative, yet unintentional event, significantly influenced evaluations of culpability: The more
research participants were led to believe a person could have foreseen a negative event, the more
responsibility and blame was assigned to them for their inaction. Several theoretical models of
blame include some form of foreseeability (Alicke, 2000; Guglielmo et al., 2009; Shaver, 1985).
Lagnado and Channon (2008) clarify conceptual distinctions in previous models and provide
empirical evidence that both subjective foreseeability (how likely an event/outcome is from an
actor's point of view) and objective foreseeability (the actual likelihood of an event/outcome)
influence judgments of blame.

32

In addition to foreseeability, Guglielmo et al. (2009) argue that viewer perceptions of
preventability—the ability to prevent a negative event from occurring—impact assessments of
blame for unintentional actions. Consistent with this assertion, Menec and Perry (1998) found
that the preventability of an array of stigmas (e.g., heart disease, drug addiction) was negatively
associated with feelings of pity. In turn, feelings of pity were positively associated with a
willingness to assist ailing individuals. Likewise, Marlow, Waller, and Wardle (2010) found that
participants made greater attributions of blame for ailments often viewed as preventable (e.g.,
chlamydia, lung cancer) relative to ailments often viewed as unpreventable (e.g., leukemia,
breast cancer). In online settings, similar relationships between forseeability, preventability, and
blame would be expected. For instance, whether viewers believe a profile owner can or cannot
prevent negative other-generated content from being shared is likely to influence their judgments
of whether the profile owner is to blame for the content being shared. Judgments of forseeability
and preventability may predict perceptions of blame, but how do people determine if an act or an
outcome was foreseeable or preventable?
This study illustrates how judgments of forseeability and preventability have the potential
to overlap with components of the folk-model of intentionality. In a broad sense, perceivers can
judge the extent to which profile owners should foresee negative content being posted to their
profiles and should prevent such content from ever being shared (e.g., change privacy settings).
Perceivers can also judge the extent to which a profile owner could have foreseen or prevented a
specific piece of other-generated content from being shared. In the latter case, knowing that the
content has been posted (awareness) and having the ability to remove the content (skill) function
as direct indicators of forseeability and preventability. An individual who knows a negative act is
occurring (awareness) can foresee specific, negative outcomes to a much greater extent than an

33

individual unaware of the act. An individual who has the opportunity to stop a negative act (due
to awareness of the act and skill to control it) is in much better position to prevent or reduce
negative outcomes than a person who has no idea the act is taking place. As such, the
components of the folk-model of intentionality can also be used by social perceivers to determine
if a negative act was intentionally allowed to transpire or persist and they can be used to infer if a
negative outcome was intentionally brought about. This capacity may obviate the need for
operationally redundant variables while simultaneously providing an understanding of the
process through which these various considerations (desires, intentions, awareness, and skill)
interrelate to affect judgments of intentionality and blame. If the answer for what makes an event
foreseeable or preventable is skill to control the event and awareness that the event is occurring
or might be occurring, then these considerations need not be made twice. Even if these
considerations do not completely overlap, it is at least important to note that perceptions of
forseeability and preventability have the potential to influence judgments of awareness, intent,
desire, and ultimately, intentionality.

34

APPENDICES

35

Appendix A
Awareness Scale
1) The profile owner knew the picture/statement was posted on his/her wall.
2) The profile owner had no idea the picture/statement was posted on his/her wall.
3) The profile owner saw the posted picture/statement on his/her wall.
4) The profile owner was clueless that the picture/statement was posted on his/her wall.
Desire Scale
1) The profile owner wants to share the picture/statement posted on his/her wall.
2) The profile owner does not wish to share the picture/statement posted on his/her wall.
3) The profile owner prefers to share the picture/statement posted on his/her wall.
4) The profile owner would not like to share the picture/statement posted on his/her wall.
Intentions Scale
1) The profile owner chose to share the picture/statement posted on his/her wall.
2) The profile owner decided to share the picture/statement posted on his/her wall.
3) The profile owner did not plan to share the picture/statement posted on his/her wall.
4) The profile owner did not intend to share the picture/statement posted on his/her wall.
Intentionality Scale
1) The profile owner deliberately shared the picture/statement posted on his/her wall.
2) The profile owner accidentally shared the picture/statement posted on his/her wall.
3) The profile owner purposefully shared the picture/statement posted on his/her wall.

36

Appendix B
Blame
1) The profile owner is not responsible for sharing the picture/statement on his/her wall.
2) The profile owner is accountable for sharing the picture/statement on his/her wall.
3) The profile owner had nothing to do with sharing the picture/statement on his/her wall.
General Perceptions of the Profile Owner
1) I view the profile owner favorably.
2) I view the profile owner negatively.
3) I view the profile owner positively.
4) I view the profile owner unfavorably.
Task Attraction of the Profile Owner
1) I would want to study for an important test with the profile owner.
2) I would want to work on a group project with the profile owner.
3) I would feel confident relying on the profile owner's notes for an important test.

37

Appendix C
Table 1.
Means and Standard Deviations

__________________________________________________________________
Comment Type

Favorable

Neutral

Control

Intentionality Statement
Intentionality Picture

4.95A (1.34)
4.35B, D (1.52)

4.19B (1.40)
4.73A, D (1.35)

3.43C (1.11)
3.58C (1.37)

Desire Statement
Desire Picture

5.63A (1.04)
5.62A (1.19)

4.57B (1.44)
5.28A (1.27)

3.13C (1.37)
4.01D (1.34)

Dependent Variable

Notes. For each dependent variable, different subscripts indicate that means significantly differ
at p <. 05. Standard deviations are in parentheses.

38

Appendix D
Figure 1.
Belief

Desire

Intention
Skill

Awareness

Intentionality
The Folk Model of Intentionality (Malle & Knobe, 1997)

39

Appendix E
Figure 2.

Sample Stimulus (Male, Positive comment, Picture)
For interpretation of the references to color in this and all other figures, the reader is referred to
the electronic version of this dissertation.

40

Appendix F
Figure 3.

Sample Stimulus (Female, Neutral Comment, Statement)

41

Appendix G
Figure 4.
Source of Information
Self
High
Control of
Dissemination
Low

Other

Lowest Warranting
Value

Moderate Warranting
Value

Low Warranting
Value

Greatest Warranting
Value

Warranting Value of Information

42

REFERENCES

43

REFERENCES
Alicke, M. D. (2000). Culpable control and the psychology of blame. Psychological Bulletin,
126, 556-574.
Alicke, M. D. (2008). Blaming badly. Journal of Cognition and Culture, 8, 179-186.
Arpan, L. M., & Roskos-Ewoldsen, D. R. (2005). Stealing thunder: An analysis of the effects of
proactive disclosure of crisis information. Public Relations Review, 31, 425-433.
Baier, A. C. (1970). Act and intent. The Journal of Philosophy, 67, 648-658.
Davis, C. (2010, July 12). Divorce, Facebook style. Philadelphia Inquirer. Retrieved July 12,
2010, from http://www.philly.com/philly/living/98220459.html?cmpid=15585797
DeAndrea, D. C., & Walther, J. B. (2010, November). Attributions for inconsistencies between
online and offline self-presentations. Paper presented at the annual meeting of the
National Communication Association conference in San Francisco, CA.
Dӧring, N. & Pӧschl, S. (2009). Nonverbal cues in mobile phone text messages: The effects of
chronemics and proxemics. In R. Ling & S. W. Campbell (Eds.), The Reconstruction of
Space and Time: Mobile Communication Practices (pp. 109-136). New Brunswick, NJ:
Transaction Publishers
Facebook (2010). Pressroom statistics. Retrieved August 13, 2010, from
http://www.facebook.com/press/info.php?statistics
Goffman, E. (1959). The presentation of self in everyday life. Garden City, NY: Doubleday.
Guglielmo, S., Monroe, A. E., & Malle, B. F. (2009). At the heart of morality lies folk
psychology. Inquiry, 52, 449-466.
Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley.
Kalman, Y. M. (2007). Silence in text-based computer mediated communication: The invisible
component. Unpublished doctoral dissertation, University of Haifa, Israel.
Jensen, J. V. (1973). Communicative functions of silence. Review of General Semantics, 30, 249267.
Johannesen, R. L. (1974). The functions of silence: A plea for communication research. Western
Journal of Communication, 38, 25-35.
Lagnado, D. A., & Channon, S. (2008). Judgments of cause and blame: The effects of
intentionality and foreseeability. Cognition, 108, 754-770.

44

Leary, M. R. (1995). Self-presentation: Impression management and interpersonal behavior.
Dubuque, IA: Brown & Benchmark Publishers.
Malle, B. F. (2004). How the mind explains behavior: Folk explanations, meaning, and social
interaction. Cambridge, MA: MIT Press.
Malle, B. F. (2006). Intentionality, morality, and their relationship in human judgment. Journal
of Cognition and Culture, 6, 61-82.
Malle, B. F. & Knobe, J. (1997). The folk concept of intentionality. Journal of Experimental
Social Psychology, 33, 101–21.
Malle, B. F. & Knobe, J. (2001). The distinction between desire and intention: A folk-conceptual
analysis. In B. F. Malle, L. J. Moses, & D. A. Baldwin (Eds.), Intentions and
Intentionality: Foundations of Social Cognition (pp. 45-67). Cambridge, MA: MIT Press.
Malle, B. F., Moses, L. J., & Baldwin, D. A. (2001). The significance of intentionality. In B. F.
Malle, L. J. Moses, & D. A. Baldwin (Eds.), Intentions and Intentionality: Foundations
of Social Cognition (pp. 1–24). Cambridge, MA: MIT Press.
Marlow, L. A. V., Waller, J., & Wardle, J. (2010). Variation in blame attributions across
different cancer types. Cancer, Epidemiology, Biomarkers, & Prevention, 19, 1799-1805.
McCroskey, J. C., & McCain, T. A. (1974). The measurement of interpersonal attraction.
Communication Monographs, 41, 261-266.
Menec, V. H., & Perry, R. P. (1998). Reactions to stigmas among Canadian students: Testing an
attribution-affect-help judgment model. The Journal of Social Psychology, 138, 443-453.
Noelle-Neumann, E. (1974). The spiral of silence: A theory of public opinion. Journal of
Communication, 24, 42-51.
Parks, M. R. (2010). Who are Facebook friends? Exploring the composition of Facebook friend
networks. Paper presented at the annual meeting of the International Communication
Association conference in Suntec City, Singapore.
Schlenker, B. R. (1975). Self-presentation: Managing the impression of when reality interferes
with self-enhancement. Journal of Personality and Social Psychology, 32, 1030-1037.
Schlenker, B. R., & Weigold, M. F. (1992). Interpersonal processes involving impression
regulation and management. Annual Review of Psychology, 43, 133-168.
Schneider, D. J., Hastorf, A. H., & Ellsworth, P. C. (1979). Person perception. Reading, MA:
Addison-Wesley

45

Shaver, K. G. (1985). The attribution of blame: Causality, responsibility, and blameworthiness.
New York: Springer-Verlag.
Shultz, T. R. & Wright, K. (1985). Concepts of negligence and intention in the assignment of
moral responsibility. Canadian Journal of Behavioral Science, 17, 97-108.
Tice, D. M, Butler, J. L., Muraven, M. B., & Stillwell, A. M. (1995). When modesty prevails:
Differential favorability of self-presentation to friends and strangers. Journal of Personality
and Social Psychology, 69, 1120–1138.
Tiersma, P. (1995). The language of silence. Rutgers Law Review, 48, 1-100.
Tufekci, Z., & Spence, K. L. (2007). Emerging gendered behavior on social network sites:
Negotiating between the pull of the social and the fear of the stalker. Paper presented at
the annual meeting of the International Communication Association, San Francisco.
Valle, K. (2008, November 16). Bosses checking up on workers via Facebook. Charlotte
Observer. Retrieved July 12, 2010, from
http://www.charlotteobserver.com/2008/11/16/354729/bosses-checking-up-onworkers.html.
Walther, J. B. (1996). Computer-mediated communication: Impersonal, interpersonal, and
hyperpersonal interaction. Communication Research, 23, 3-43.
Walther, J. B., Carr, C., Choi, S., DeAndrea, D.C., Kim, J., Tong, S. T., & Van Der Heide, B.
(2010). Interaction of interpersonal, peer, and media influence sources online: A
research agenda for technology convergence. In Z. Papacharissi (Ed.), The networked self:
Identity, community and culture on social network sites. New York: Routledge.
Walther, J. B., & Parks, M. R. (2002). Cues filtered out, cues filtered in: Computer-mediated
communication and relationships. In M. L. Knapp & J. A. Daly (Eds.), Handbook of
interpersonal communication (3rd ed., pp. 529-563). Thousand Oaks, CA: Sage.
Walther, J. B., & Tidwell, L. C. (1995). Nonverbal cues in computer-mediated communication,
and the effect of chronemics on relational communication. Journal of Organizational
Computing, 5, 355-378.
Walther, J. B., Van Der Heide, B., Hamel, L., & Shulman, H. (2009). Self-generated versus
other-generated statements and impressions in computer-mediated communication: A test
of warranting theory using Facebook. Communication Research, 36, 229–253.
Walther, J. B., Van Der Heide, B., Kim, S., Westerman, D., & Tong, S. T. (2008). The role of
friends’ behavior on evaluations of individuals’ Facebook profiles: Are we known by the
company we keep? Human Communication Research, 34, 28–49.

46

Westerman, D., Van Der Heide, B., Klein, K. A., & Walther, J. B. (2008). How do people really
seek information about others?: Information seeking across Internet and traditional
communication channels. Journal of Computer-Mediated Communication, 13, 751-767.
Williams, K. D., Bourgeois, M. J., & Croyle, R. T. (1993). The effects of stealing thunder in
criminal and civil trials. Law and Human Behavior, 17, 597–609.

47